In recent years, society has been increasingly dependent on IT systems as the IT systems have been applied to more areas. As a result, the IT systems have been turned into a social infrastructure. As the international community has been globalized, data centers that make up the IT systems are spread across the world. In response to such circumstances, the operational control and failure response operations of the IT systems are conducted in more complicated ways and on larger scale. Given the above fact, it is necessary to carry out investigation into failures on the basis of the assumption that the operation of a system is complex as well as to collect information in order to carry out investigation into failures.
As techniques of collecting information needed to carry out investigation into failures as described above, a collection tool and a collection agent are well known. The following describes the techniques with reference to the accompanying drawings. FIG. 18 is a block diagram illustrating a conventional information collection system.
As illustrated in FIG. 18, the conventional information collection system includes servers 901 to 903 and a collection tool 801 connected to the servers through a network. The servers 901 to 903 are equipped with agents 904 that collect information. The agents 904 collect information pertaining to predetermined items at predetermined intervals or at a predetermined time and transmit the collected information to the collection tool 801. Thanks to the above configuration, the conventional IT system uses the agents 904 installed in the servers to collect information and uses the information that the collection tool 801 acquires from the agents 904 to carry out investigation into failures.
What is known as a conventional technique related to the present invention is a data collection device, a data collection system and a data collection method that: calculates a usage amount rate of each factor for each to-be-collected data item; generates a schedule that enables data to be collected at a time when a calculated operating rate is low; and collects the predetermined data on the basis of the generated schedule for carrying out collection (see Patent Document 1, for example). [Patent Document 1] Japanese Laid-open Patent Publication No. 2006-79488
However, the problem with the information collection method of the conventional IT system is that due to the delay of network communication and different loads of information collection operations on the agents 904 of the servers 901 to 903, a time lag occurs between the pieces of information acquired by the collection tool 801 as a result. Another problem is that since the time lag occurs between the acquired pieces of information, it is difficult to grasp the causal relationship between the pieces of information and, as circumstantial evidence, the reliability is low. Moreover, according to the information collection technique of the conventional IT system, information is collected at predetermined intervals or at a predetermined time. Therefore, for example, the problem is that it is impossible to collect information immediately after a failure occurs or at other arbitrary timings.